University of Groningen
Autonomy and Chirality in Molecular Motors
Kistemaker, Jozef Cornelis Maria
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2017
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Kistemaker, J. C. M. (2017). Autonomy and Chirality in Molecular Motors. Rijksuniversiteit Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
119 This chapter has been published as:
J. C. M. Kistemaker, S. F. Pizzolato, T. van Leeuwen, T. C. Pijper, B. L. Feringa, Chem. - Eur. J. 2016, 22, doi:10.1002/chem.201602276.
Chapter 5: Identification of Two Thermal Isomerization
Pathways for Bistable Molecular Motors
Herein is reported: Chiroptical molecular switches play an important role in responsive materials and dynamic molecular systems. Here we present the synthesis of four chiral overcrowded alkenes and the experimental and computational study of their photochemical and thermal behaviour. By irradiation with UV light, metastable diastereoisomers with opposite helicity were generated through high yielding E–Z isomerizations. Kinetic studies on metastable 1–4 using CD spectroscopy and HPLC analysis revealed two pathways at higher temperatures for the thermal isomerization, namely a thermal E–Z isomerization (TEZ) and a thermal helix inversion (THI). These processes were also studied computationally whereby a new strategy was developed for calculating the TEZ barrier for second generation overcrowded alkenes. In order to demonstrate that these overcrowded alkenes can be employed as bistable switches, photochromic cycling was performed, which showed that the alkenes display good selectivity and fatigue resistance over multiple irradiation cycles. In particular, switch 3 displayed the best performance in forward and backward photoswitching, while 1 excelled in thermal stability of the photogenerated metastable form. Overall, the alkenes studied showed a remarkable and unprecedented combination of switching properties including dynamic helicity, reversibility, selectivity, fatigue resistance and thermal stability.
120
5
Introduction
Responsive materials[1–3] and dynamic molecular systems[4–10] that change structure and functions as a result of an external input signal are attracting major attention owing to the prospect of smart materials[11–17] and nanoscale devices.[18–22] Photochemical switches allow for non-invasive control, reversibility and high spatio-temporal precision.[5] Overcrowded alkenes have been used in a wide variety of responsive nanoscale systems, such as in a molecular car powered by four unidirectional molecular motors or as multi-state switches featuring dynamic functions of which up to four distinct configurations can be addressed.[23–25] The large number of structural modifications that has been presented by our group and others has expanded the field of molecular design with three generations of photoswitchable overcrowded alkenes, the majority of which exhibits a strong directional preference and functions as rotary motors.[7,26–32] Through desymmetrization of our systems, the unidirectionality of the rotary motion has been extensively demonstrated and various stereoisomers have been identified by spectroscopic and chromatographic techniques for each variation in design.[33–44] A key aspect of these systems is that the photochemical generation of metastable species is followed by thermally induced isomerizations, for which the life-times have been tuned through structural changes to range from nanoseconds to years.[34,38,44,45] Hence, overcrowded alkenes can be defined as either motors or switches depending on the activation energy and therefore speed of the thermal isomerization step (i.e. when the rotation rate is the limiting step). Their propensity to undergo continuous light- and thermal-induced directional rotary motion (motor behaviour), however, diminishes their usefulness as switches in applications where thermal stability is desired, for example, in the field of photoswitchable catalysis.[24,46,47] As such, there is a demand for thermally highly stable alkenes that can be switched photochemically and reversibly between distinct geometrical chiral forms.
Scheme 5.1. General scheme of photochemical E–Z isomerization and thermal helix inversion of
second generation molecular motors.
Molecular motors of the second generation consist of a symmetric ‘lower’ half (for R=H) and an asymmetric ‘upper’ half that feature a single stereocentre (Scheme
121
5
BISTABLE MOTORS
5.1).[30,48] Upon irradiation with UV light they can undergo a photochemical E–Z isomerization, a process that results in a metastable (MS) diastereoisomer which is of the opposite helicity. In this process, the methyl substituent on the stereogenic centre has changed from an unhindered outward facing axial orientation to an equatorial orientation in which the methyl faces the lower half, thus creating steric hindrance. This steric strain causes the MS diastereoisomer to be higher in energy with respect to the original configuration. The strain can subsequently be reduced by a thermally activated isomerization in which (usually) the upper half moves around the lower half, again resulting in an inversion of the helicity. In the resulting stable isomer, the upper half has undergone a 180° rotation with respect to the lower half (see Scheme 5.1, in the case R = H, the symmetry in the lower half causes the initial and final states to be chemically identical). In theory, it is possible that the thermal isomerization of the metastable state follows an alternative and competing pathway other than thermal helix inversion (THI). Structurally similar stilbene switches are able to undergo thermal E–Z isomerization (TEZ) from cis to trans, although the activation energy for this process usually exceeds 150 kJ·mol−1.[49,50] For some overcrowded alkenes though, this barrier has been observed to be significantly lower due to the steric strain in the minimum energy configurations, thus forcing the double bond far from planarity. As an example, bis-fluorenylidenes exhibit activation energies for the TEZ of ~105 kJ·mol−1.[51,52] For second generation motors, in order to positively identify the outcome of the thermal isomerization of the photochemically generated metastable state the lower half has to be desymmetrized. The two step process starting from the stable-(Z) state will then lead either to the opposite isomer stable-(E) of the initial configuration, which is indicative of a THI, or back to the initial stable isomer stable-(Z), thus indicating a reversible switching process by a TEZ (Scheme 5.2).
Previously, research on molecular motors has to a large extent been supported by computational chemistry.[31,48,53–55] For example, it has been shown that the energy barrier of the thermal helix inversion can be predicted with reasonable accuracy (within several kJ·mol−1) through the use of density functional theory (DFT) at the B3LYP/6-31G(d,p) level.[31,44,48,53,55] Indeed, this method has been utilized to design new motors in silico by prediction of the helix inversion energy barriers of motors prior to their synthesis to determine whether their rotation rates would be of the desired order of magnitude. However, to date a method has not been available that can accurately predict the energy barrier of the potentially competing TEZ. The reason for the lack of a suitable computational method for investigating the TEZ is that the transition state (TS) involved cannot be accurately described by single determinant methods such as Hartree−Fock and Kohn−Sham DFT. For this reason,
122
5
a multideterminant method has to be employed instead. One popular multideterminant method potentially capable of providing an accurate description is the CASSCF theory in which the orbitals of interest are treated as in a full configuration interaction (CI) calculation (the active space).[56] Such an approach can be very effective in describing the ‘static correlation’ that often cannot be accounted for effectively by a single determinant method. Unfortunately, the use of CASSCF is limited by the number of orbitals that can be treated in this way, since full CI calculations are computationally demanding. In practice, the size of a complete active space is often limited to 14 electrons in 14 orbitals to keep the calculation manageable.[57] This poses a challenge as the π system of a molecular motor is typically much larger. As such, if CASSCF or a related theory is to be used for studies on molecular motors, it is necessary to limit the number of CSFs in such a way that the accuracy is not significantly impaired.
Scheme 5.2. General scheme for photochemical and thermal behaviour (TEZ vs. THI) of
desymmetrized overcrowded alkenes stable-(Z) and metastable-(E).
Herein is reported on the switching behaviour of four second-generation overcrowded alkenes, namely 1–4 (Scheme 5.3). Their photochemical and thermal isomerization processes have been studied by various analytical methods, while the thermal isomerization processes are also studied by computational methods. A strategy for studying the TEZ by computational chemistry will also be presented. We will show that the metastable isomers of 1–4 are able to undergo thermal isomerization through both the THI and TEZ pathways. Finally, we will demonstrate that 1–4 exhibit properties that make them highly useful bistable switches, such as high selectivity, low switching fatigue, and high thermal stability.
123
5
BISTABLE MOTORS
Results and Discussion
Design
As mentioned above, the bridging units (X and Y, see Scheme 5.1) included in the rings connected by the tetrasubstituted alkene play an important role in the structure’s flexibility, thermal stability, and switching properties. Previous studies on overcrowded alkenes with symmetrical lower halves have shown the effect of the size of the rings connected to the bridging alkene bond on the activation barrier of the thermal relaxation step.[48] In particular, the combination of a five-membered ring in the lower half (fluorene) with a sulfur or oxygen containing six-membered ring in the upper half (Scheme 5.1, benzo[f]thiochromene (X=S, Y=−) and benzo[f]chromene (X=O, Y=−), respectively) resulted in distinctive high energy activation barriers for the thermal relaxation step and consequently long half-lives of the metastable species (Δ‡G° = 109 kJ·mol−1, t
½ at room temperature (rt) = 35 days (X=S) and Δ‡G° = 106 kJ·mol−1, t
½ at rt = 9.4 days (X=O), respectively). Due to the lack of asymmetry in the lower half, the two aforementioned competing thermal pathways, that is, THI and TEZ, could not be distinguished, as they would give access to the two undistinguishable products. Therefore, we decided to extend our investigation to four overcrowded alkenes of the second generation with an asymmetric substitution pattern in the lower half (1–4, Scheme 5.3), expecting these systems to display thermal bistability. The overcrowded alkenes were synthesized by Stefano Pizzolato and Tom van Leeuwen following procedures similar to those used for analogous systems.[29,34,48] The enantiomers of 1–4 were separated by preparative chiral HPLC (Figure 5.1, Figure 5.2 and ref. [58] for full details).
Scheme 5.3. Second generation overcrowded alkenes Photochemical and thermal isomerizations
Single enantiomers of each overcrowded alkene were subjected to circular dichroism (CD) spectroscopy to assign absolute stereochemistry as well as to perform a qualitative analysis. The isolated enantiomers of compounds 1–4 displayed strong Cotton effects in the area of ~250–320 nm and slightly smaller Cotton effects of opposite sign at higher wavelengths (>320 nm) with the exception
124
5
of compound 3 which lacked such a longer wavelength absorption band (Figure 5.1). The presence of such strong Cotton effects around 400 nm is indicative of the helical shape of these overcrowded alkenes, while the lack of this band for compound 3 could be due to the absence of a heteroatom in its core structure (which is present in the other compounds).
Figure 5.1. CD spectra of 1–4. Black: experimental CD spectra of (Z,P,S)-1, (Z,P,S)-2, (Z,P,R)-3,
and (Z,M,R)-4 (heptane, 1.0·10−5 M). Blue: theoretical ECD spectra calculated with TD-DFT,
normalized and shifted by 30 nm (B3LYP/6-31+G(d,p) applying Gaussian shapes (line width = 0.3 eV) to 30 discrete transitions) used to assign enantiomers. Red: CD spectra of PSS mixtures of 1 (312 nm), 2 (365 nm), 3 (312 nm), and 4 (365 nm). Irradiation conditions: the PSS mixtures were obtained starting from the above mentioned solutions (heptane, 1.0·10−5 M, black curves)
after irradiation (indicated wavelength) at rt over 2 min.
To assign absolute stereochemistry, experimentally obtained CD spectra were compared to the calculated CD spectra. Potential energy surfaces of 1–4 were investigated with the semi-empirical PM3 method and the geometries of the resulting minima and transition states were refined by using DFT (B3LYP/6-31G(d,p) (vide infra). Time-dependent (TD) DFT with the B3LYP functional and a 6-31+G(d,p) basis set provided theoretical CD spectra of 1–4 and allowed the assignment of the absolute stereochemistries of 1–4 (Figure 5.1). Due to the existence of multiple conformations (e.g. of the methoxy group) and the uncertainty in the calculated Boltzmann distribution used to proportionate the spectra of the individual conformations, the calculated spectra were not expected to display a complete match with the experimental data. However, the match is sufficient to allow the discrimination between the two possible enantiomers and is therefore
125
5
BISTABLE MOTORS
suitable for the assignment of the absolute stereochemistry of 1–4.[31,40,55,59] The chiral descriptors for each species described in this chapter (e.g. (Z,P,S)-1, Figure 5.1) indicate, respectively: the configurational isomer of the tetrasubstituted alkene (E or Z), the configurational helicity of the molecule (P or M), and the absolute stereochemistry of the stereogenic centre (R or S).
Figure 5.2. Chiral HPLC traces of 1–4. Top: HPLC traces (heptane/2-propanol) of pure
enantiomers separated by preparative chiral HPLC (structures of stable-Z isomers depicted in Figure 5.1) as assigned by CD absorption spectroscopy: 1 (Chiralcel OD-H, 98:2),
(Z,P,S)-2 (Chiralpak AD-H, 97:3), (Z,P,R)-3 (Chiralcel OD-H, 98:(Z,P,S)-2), and (Z,M,R)-4 (Chiralcel OD, 99.3:0.7).
Middle: HPLC traces of the PSS mixtures of 1–4 (identical conditions). Bottom: HPLC traces after subsequent thermal isomerization of 1–4 (identical conditions). For irradiation and thermal isomerization conditions, see Figure 5.1 and Figure 5.4. MS=metastable.
The correlation between the Cotton effect and the helicity agrees with the results of Cnossen et al. in which the same correlation was observed for four different overcrowded alkenes.[55] Compounds with a positive helicity display a negative Cotton effect for the longest wavelength absorption band and vice versa, with the exception of 3 as this species lacks a strong CD absorption band in the 350–450 nm region (vide supra). UV irradiation of solutions in heptane (312 or 365 nm, see Figure 5.1 for details) of each of the Z isomers of 1–4 resulted in the inversion of the major bands in their CD spectra. This is indicative of an inversion in helicity and shows that the photochemical Z-E isomerization of the stable-(Z)-1–4 to the metastable-(E)-1–4 has taken place. The presence of the metastable-(E) isomers was further confirmed by chiral HPLC analysis (Figure 5.2) and 1H NMR spectroscopy (illustrated for 1 in Figure 5.3, see ref. [58] for further details). The
126
5
ratio between the E and Z isomers at the photostationary state (PSS) in heptane solution was determined by chiral HPLC analysis (E:Z ratio: (S)-1 95:5, (S)-2 96:4, (R)-3 97:3, and (R)-4 99:1), showing an almost quantitative photoswitching process towards the metastable diastereoisomer for all four compounds, with remarkably high ratios for this class of overcrowded alkene based switches.
Figure 5.3. 1H NMR spectra of the switching process of 1. a): 1H NMR spectra of stable-(Z)-1
(~3 mg in CDCl3, 0.8 mL); b): 1H NMR spectra after irradiation of stable-(Z)-1 (312 nm) to the
metastable state affording a PSS mixture in CDCl3 of stable-(Z)-1 : metastable-(E)-1 = 16 : 84 (note:
PSS ratios are known to be affected by the nature of the solvent, vide supra).
Heating the irradiated samples allowed them to undergo thermal isomerization (for conditions, see Figure 5.4), which resulted in major changes in their CD spectra. HPLC chromatograms of the resulting samples showed a partial reversal of the metastable-(E) isomer to the initial, stable-(Z) isomer which is observed together with the appearance of a new peak attributed to the stable-(E) isomer (Figure 5.2). These results signify: i) that the photochemical Z to E isomerization results in the formation of a highly stable diastereoisomer that is able to relax measurably only at high temperatures, and ii) that relaxation can take place via two competing pathways, one leading to the initial isomer through TEZ and the other leading to the corresponding E isomer through THI (Scheme 5.2).
To investigate the kinetic behaviour of the two thermal isomerization pathways, samples of alkenes 1–4 were irradiated to PSS at room temperature after which their thermal relaxation was monitored over time. For alkenes 3 and 4, thermal relaxation was followed in real-time using CD spectroscopy at the specific wavelength (381 nm and 395 nm, respectively) which showed the largest difference between the initial stable-(Z) isomer and the PSS mixture (Figure 5.4). However, for alkenes 1 and 2 this setup was not suitable as the thermal relaxation
127
5
BISTABLE MOTORS
of these alkenes, in order to become observable, required temperatures that are above the temperature range of the temperature controller of the CD spectrophotometer employed. Instead, solutions of (Z,P,S)-1 and (Z,P,S)-2 in hexanol and dodecane, respectively, were irradiated to PSS and placed in a temperature controlled oil bath. Aliquots were then taken regularly and analysed by chiral HPLC. A least squares analysis of the HPLC integrals of the major diastereoisomers of 1 and 2 versus time and the change in CD absorption for 3 and 4 versus time provided the reaction rates (ktotal) for the thermal isomerization process at various temperatures. The observed rate is the sum of the individual rates for TEZ and THI (kTEZ and kTHI) and these are related as given in Equation 1:
1
where the final ratio between the stable-(Z) and stable-(E) isomers is obtained from HPLC after correction for the initial concentration of the stable-(Z) isomer at PSS. A least squares analysis of the rates versus the temperature on the original Eyring equation:
°
‡ ‡ °
2
with appropriate weighing (1/k2) afforded the entropies and enthalpies of activation. The standard errors (σ) were obtained from a Monte Carlo error analysis on the linearized Eyring equation:
∙ ln ∙ ln ‡° ‡ ° 3
from forty thousand samples using calculated standard errors on rates and estimated standard errors on temperatures. The results of the Eyring analysis are summarized in Table 5.1. From the fitting curves in Figure 5.4 it is evident that the increase in temperature is accompanied by a decrease in accuracy. This is expected for these experiments and therefore an extensive error analysis has been performed to assure the validity of the results from the Eyring analysis. While the error on the derived enthalpy (Δ‡H°) and entropy (Δ‡S°) of activation are appreciable, the error on the Gibbs free energy of activation (Δ‡G) remains small, particularly when it is calculated for a temperature in or near the range of temperatures in which the thermal relaxation was observed. The reason for this is that extrapolation of these parameters to room temperature spans over hundred degrees Celsius for some examples thus magnifying the uncertainty. This is notably observed for the half-life at room temperature, an often reported feature.
128
5
Figure 5.4. Decay curves (top/left axes) and Eyring plots (bottom/right axes) of metastable 1–4.
Decay curves of: (E,M,S)-1 recorded by HPLC taking aliquots from a hexanol solution (131– 152 °C); (E,M,S)-2 recorded by HPLC taking aliquots from a dodecane solution (112–132 °C); (E,M,R)-3 recorded by CD in dodecane (95–105 °C); (E,P,R)-4 recorded by CD in dodecane (84– 105 °C). Least squares analysis on the original Eyring equation for 1–4 with error bars of 3σ. (heptane, 1.0·10−5 M). Thermal decay conditions: the PSS mixtures (hexanol or dodecane,
1.0·10−5 M) were heated at fixed temperatures starting from the above mentioned solutions (black
curves) after irradiation with UV light (indicated wavelength) at rt over 2 min under stirring.
Table 5.1. Kinetic parameters determined by the direct Eyring analysis (Figure 5.4), with standard
errors obtained from a Monte Carlo analysis for thermal isomerizations (TEZ and THI) of MS-1–4.
(E,M,S)-1 (E,M,S)-2 (E,M,R)-3 (E,P,R)-4
t½ at rt (years)[a] 75±35 4.3±4.2·104 1.3 ±0.6 1.3±0.2 T at t½=1 h (°C) 138.2±0.4 121.3±0.3 116.0 ±0.7 99.1±0.2 Δ‡H° TEZ (kJ·mol−1) 110±5.4 184±8.1 86.6 ±4.1 108±2.6 Δ‡S°TEZ (J·K−1·mol−1) −51.3±13 147±21 −98.5 ±11 −30.9±7.1 Δ‡G°TEZ (kJ·mol−1)[b] 129±0.6 129±0.5 123 ±0.1 120±0.5 Δ‡H° THI (kJ·mol−1) 118±5.7 208±9.2 99.4 ±4.6 96.5±2.4 Δ‡S°THI (J·K−1·mol−1) −53.2±14 182±23 −72.8 ±12 −68.0±6.6 Δ‡G°THI (kJ·mol−1)[b] 138±0.6 140±0.6 127 ±0.1 122±0.5 [a] rt: 20 °C.
[b] Standard condition: 100 °C and atmospheric pressure.
For example, the standard error determined for the half-life of 2 is as large as the half-life itself. Therefore, we report a more appropriate characteristic of the first order reaction, namely the one-hour-half-life temperature which is the temperature at which the half-life equals one hour. This property is not an extrapolation but usually falls within or close to the range of measured temperatures and is derived
129
5
BISTABLE MOTORS
from the Eyring equation by the use of the Lambert W function[60] as in Equation 4: / ° ‡ ∙ ∙ ∙ ∙ ‡ °∙ ° ‡ ∙ ∙ 4
The error on the parameters discussed for different processes is reduced to less than a percent of the parameter. Moreover, the temperature at which the half-life equals one hour is a much more chemically intuitive feature, particularly when the half-lives at room temperature of the processes under investigation are over a year or even exceeding forty thousand years (as for (E,M,S)-2). Going from oxygen in 4 (X=O), carbon in 3 (X=C), to sulfur in 2 (X=S), the one-hour-half-life temperature increases from 99 to 116 and 121 °C, indicating an increase in stability of the metastable diastereoisomer. Furthermore, substituting the naphthalene moiety in 2 for the xylene moiety in 1 increases the one-hour-half-life temperature even further to 138 °C.
Figure 5.5. Gibbs free energy of activation for the TEZ and the THI processes plotted versus
temperature for the metastable diastereoisomers of alkenes 1–4. The experimental temperature range is marked by grey vertical lines (85–152 °C), inversion points for the two processes are marked for 3 and 4 by a dot.
130
5
Figure 5.6. Calculated geometries of (S)-2 (X = S). Left: top view with the upper half on top, the
alkene on the z-axis and the fluorene in the x-z plane, right: front view with the alkene on the y-axis and the fluorene in the x-y plane.
131
5
BISTABLE MOTORS
From the Gibbs free energy of activation for the two possible pathways it is clear that under standard conditions the TEZ pathway is preferred over the THI pathway. Plotting the Gibbs free energy versus temperature (Figure 5.5), hereby assuming that the enthalpy and entropy are temperature independent, reveals that for the entire temperature range under investigation (experimental temperature range: 85– 152 °C) the barrier for the TEZ is lower than that for the THI for all alkenes 1–4. A difference in the entropy of activation for the two processes logically leads to the existence of a point at which the rates for the two processes are expected to be equal. Such a point would signify the inversion of the two processes, because beyond this temperature the barrier of the THI will be lower than that of the TEZ. For alkenes 3 and 4, these points are found at relevant temperatures (37.8 °C and 226 °C, respectively) while for 1 and 2 the inversions would take place either far outside of the experimentally significant temperature range or never at all (411 °C and <0 K, respectively).
Computational results
The experimental study of the thermal behaviour of the metastable diastereoisomers (E)-1–4 was accompanied by a computational study of the potential energy surface of overcrowded alkenes 1–4. As indicated above, a semi-empirical scan of the PES (PM3) provided minima and transition states which were refined by using DFT with the B3LYP functional and a 6-31G(d,p) basis set.[61–68] However, while Kohn–Sham DFT theory is adequate for studies into the THI pathway, studies into the TEZ pathway require the use of a method capable of accounting for static correlation. The reason for this is that the TS of the TEZ has strong biradicaloid character which cannot be properly described with a single-determinant method.[53] As such, calculations on the TEZ pathway were performed by using CASSCF and MRMP2 theories. The TEZ pathway calculations were performed by Thom Pijper.[69]
Initial TEZ TS geometries were obtained at the GVB-PP(1) level (which gives results identical to CASSCF(2,2)) using the 3-21G basis set. The initial geometries were subsequently refined at the CASSCF(10,10)/6-31G(d) level. In cases where more than one transition state geometry was found, subsequent calculations were performed for all geometries to determine their energies. Each optimized geometry was subjected to a vibrational analysis to verify whether it truly corresponds to a minimum on the PES (or first-order saddle point in the case of a TS geometry). From these calculations, thermochemical corrections (at T = 373.15 K and p = 1 atm) to the calculated energies were also obtained. The CASSCF(10,10)/6-31G(d) optimized geometries were finally used to calculate energies at the
132
5
MRMP2/CASSCF(14,14)/6-31G(d) level (Figure 5.6), to which the thermochemical corrections were subsequently added. It should hereby be noted that energy calculations at the MRMP2/CASSCF(14,14)/cc-pVTZ level were attempted as well, but were found to display a larger deviation from experimental results. This observation, combined with the use of an active space which excludes part of the π system as well as lone pairs of heteroatoms, suggests that the predictive accuracy of the presented method is in part due to a favourable cancellation of errors. A more detailed discussion of the computational approach used is presented in ref. [58].
Table 5.2. Relative Gibbs free energies of 1–4 calculated at the DFT-B3LYP/6-31G(d,p) level or
MRMP2/CASSCF(14,14)/6-31G(d) // CASSCF(10,10)/6-31G(d) level indicated by * (373.15 K, 1 atm, in kJ·mol−1). (S)-1 (S)-2 (R)-3 (R)-4 Stable-(Z) 0 0 0 0 TS TEZ* 165 150 146 138 Metastable-(E) 22.2 23.1 27.2 24.9 TS THI 163 165 151 146 Stable-(E) 0.94 1.74 1.89 1.46
MS-(E) X-ring size (pm)[a] 961 958 901 875
MS-(E) dihedral angle (°)[b] 46.7 47.6 42.6 41.4
[a] Summed lengths of the bonds making up the six membered ring in the upper half (Figure 5.7b). [b] Dihedral angle made up by atoms 1, 2 and the central alkene as indicated in Figure 5.7a.
Figure 5.7. a) Active space used in CASSCF(14,14) and MRMP2/CASSCF (14,14) calculations. π
bonds included are indicated with red circles. b) Correlation between X-ring size and thermal relaxation energy barrier.
The calculated Gibbs free energies are summarized for alkenes 1–4 in Table 5.2 and the obtained geometries for (S)-2 are depicted in Figure 5.6 as an example. The geometries of 1, 3, and 4 do not differ significantly in general appearance from those of 2, although they naturally do differ in specific bond angles and lengths. Going from alkene 4 to alkene 3 to alkene 2, the size of the bridging atom X in the upper half increases, which is accompanied by an increase in the size of that ring and thus forces the aryl moiety towards the lower half (Figure 5.7b, Table 5.2). The increase in steric hindrance is alleviated by additional folding of the six membered
133
5
BISTABLE MOTORS
ring, as can be seen from the dihedral angle made up by atoms 1, 2, and the central alkene (as indicated in Figure 5.7a for the metastable-(E) isomer). Both 1 and 2 are bridged by a sulfur atom and therefore hardly differ in ring size, although the difference between the aryl and xylyl moieties is to a small extent reflected by their dihedral angles (see ‘dihedral angle’ in Table 5.2). The barrier for the THI increases with an increase in the degree of folding of the upper half, as is seen from the dihedral angle. A similar increase is observed for the calculated barrier for the TEZ, with the exception of 1 which exhibits a significantly higher barrier without an increase in folding with respect to 2. A rationale for this difference in activation energy is that the six-electron π-system of the xylyl moiety of 1 is less effective in stabilizing the biradicaloid character of the TS geometry compared to the larger 10-electron π-system of the naphthyl moiety of 2.[58]
Table 5.3. Comparison of experimental and theoretical barriers for TEZ and THI of 1–4.[a]
Metastable: (E)-1 (E)-2 (E)-3 (E)-4
Δ‡G°TEZ (kJ·mol−1) 129±0.6 129±0.5 123±0.1 120±0.5
Δ‡GcalcTEZ (kJ·mol−1) 142 127 119 113
Δ‡G°
THI (kJ·mol−1) 138±0.6 140±0.6 127±0.1 122±0.5
Δ‡GcalcTHI (kJ·mol−1) 141 142 124 121
[a] standard condition: 100 °C and atmospheric pressure.
Table 5.3 provides an overview of the experimentally determined and calculated Gibbs free energies of activation for the TEZ and THI pathways for alkenes 1–4. The calculated barriers for the THI agree strongly with those found experimentally, differing by no more than 3 kJ·mol−1. The calculated barriers for the TEZ deviate more from the experimentally determined barriers. The barriers of 2 and 3 correspond well whereas the barrier of 1 is overestimated and the barrier of 4 is slightly underestimated. The slight underestimation of the TEZ barrier for 4 still allows for a reasonable prediction of the behaviour of the overcrowded alkene, however, the overestimated barrier of 1 suggests almost equal rates for the THI and TEZ processes while experimental results show the TEZ pathway to be significantly faster than the THI pathway. This could imply that the computational approach used herein may not be as accurate for overcrowded alkenes with xylene-derived upper halves as it is for those with naphthalene-xylene-derived upper halves. Nonetheless, these computational methods provide valuable insight into how the thermal isomerization behaviour relates to the geometric changes in these second generation overcrowded alkenes.
134
5
Photoswitching process
It was found that the increased thermal stability of the metastable states of alkenes 1–4 makes them very suitable candidates for use as bistable photoisomerizable switches. The switching properties of 1–4 were monitored by UV-vis absorption spectroscopy (Figure 5.8) and 1H NMR spectroscopy (vide supra). Solutions of stable 1–4 (heptane) in quartz cuvettes were irradiated at room temperature for a few minutes towards either the metastable state using UV light (312 or 365 nm) or the stable state using visible light (420 or 450 nm). Using UV-vis absorption spectroscopy, the reversible photochemical E–Z isomerizations were found to be characterized by clear isosbestic points, indicating the absence of side reactions, as well as bathochromic shifts of the major absorption bands in the metastable state of about 30–80 nm.
Figure 5.8. UV-vis spectra of the switching process of 1–4. Experimental UV-vis absorption spectra
in black of (Z,P,S)-1, (Z,P,S)-2, (Z,P,R)-3, (Z,M,R)-4 (heptane, 1.0·10−5 M). Irradiation of 1
(312 nm), 2 (365 nm), 3 (312 nm), and 4 (365 nm) to the metastable state affords a PSS shown in red with two intermediate moments in the process shown as well in red (E:Z ratio: 1 95:5,
(S)-2 96:4, (R)-3 97:3, and (R)-4 99:1). Irradiation using a longer wavelength of 1 (4(S)-20 nm), (S)-2 (450 nm), 3 (420 nm), and 4 (450 nm) allowed for the reversed E–Z isomerization towards the stable state
affording a new PSS shown in blue with two intermediate moments in the process shown as well in blue (Z:E ratio; (S)-1 64:36, (S)-2 82:18, (R)-3 97:3, (R)-4 70:30). Inserts display irradiation cycles between the two PSS’s for each compound.
This is in full agreement with the calculated structural change and concomitant change in the HOMO-LUMO gap. Upon the formation of the metastable state the overcrowded alkene experiences an increase in twist as is illustrated in the calculated structures in Figure 5.6. This twisting increases the energy of the HOMO
135
5
BISTABLE MOTORS
by an average 23 eV for 1–4 while at the same time lowering the LUMO by an average 246 eV, which together significantly reduces the HOMO-LUMO gap. This is opposite to the observations of Cnossen et al. for second generation molecular motors with six membered rings in both the upper as well as lower half in which the twist over the double bond was lowered in the metastable state and a hypsochromic shift was observed.[55] The bathochromic shift upon formation of the metastable 1–4 allows for a highly selective photochemical switching process in which both states can be addressed by the use of light of an appropriate wavelength (Figure 5.8, Scheme 5.4). None of the overcrowded alkenes showed any noticeable degradation over multiple switching cycles, thus exhibiting an excellent fatigue resistance of this family of molecular switches.
Scheme 5.4. General scheme for reversible highly selective photoswitching of stable (Z)-1–4 and
metastable (E)-1–4.
The PSS ratios for the stable-(Z) to metastable-(E) isomerizations obtained upon irradiation with UV light were determined by HPLC and were all found to yield ≥ 95% of the metastable-(E) state for the forward isomerization (vide supra). The reverse reaction using visible light afforded varying PSS ratios (Z:E ratio; (S)-1 64:36, (S)-2 82:18, (R)-3 97:3, (R)-4 70:30), as determined by HPLC and/or UV-vis (see ref. [58]). Despite considerable overlap in the UV, the PSS towards metastable is nearly quantitative, while the reverse PSS towards stable is not, indicating that stable to metastable quantum yields are much higher than metastable to stable quantum yields. This has also been observed for a series of second-generation molecular motors.[55] Alkene 3 hereby displayed the most efficient photoswitching, producing 97% of the opposite diastereoisomer in both directions, and would therefore be the most suitable candidate for use as a bistable photocontrolled switch. With respect to thermal stability, overcrowded alkene 1 exhibits the most favourable behaviour, possessing a one-hour-half-life temperature of 138 °C. This is over 22 °C higher than that of 3 and the TEZ of 1 yields the starting isomer almost exclusively (>94%), making it remarkably bistable as well as selective during the thermal isomerization. Additional studies regarding the behaviour of these motors starting from their stable-(E) isomer have been reported by Stefano Pizzolato.[70,71]
136
5
Conclusion
Four overcrowded alkenes have been synthesized and investigated experimentally and computationally. The calculated CD spectra of 1–4 agree well with the experimental spectra which allowed their absolute stereochemical assignments. Irradiation with UV light allowed high yielding E–Z isomerizations providing metastable diastereoisomers. Kinetic studies on metastable 1–4 using CD and HPLC identified two pathways at high temperatures for thermal isomerization. The thermal E–Z isomerizations and helix inversions were studied computationally and a new strategy was developed for calculating the TEZ barrier for second generation overcrowded alkenes. By including the most important π bonds in the active space, the method is able to provide energy barriers for the TEZ process that are in good agreement with those observed experimentally. Furthermore, the calculated THI barriers were found to be in close agreement with those observed experimentally. To show the value of these overcrowded alkenes as bistable switches, photochemical switching cycles were performed which proved the alkenes to be excellent switches. Switch 3 showed the best performance as a photo-switch, while 1 excelled in thermal stability, both exhibiting highly selective isomerizations. These favourable switching properties offer attractive prospects for the design of novel photoresponsive systems.
Acknowledgements
This work was executed in collaboration with Stefano Pizzolato. Motors 1–3 have been synthesized by Stefano Pizzolato and motor 4 has been synthesized by Tom van Leeuwen.[71] The computational study of the TEZ pathway has been performed by Thom Pijper.[69]
Experimental Section
General Methods
Chemicals were purchased from Sigma Aldrich, Acros or TCI Europe N.V. Solvents were reagent grade and distilled and dried before use according to standard procedures. Dichloromethane and toluene were used from the solvent purification system using an MBraun SPS-800 column. Tetrahydrofuran was distilled over sodium under a nitrogen atmosphere prior to use. Column chromatography was performed on silica gel (Silica Flash P60, 230–400 mesh).NMR spectra were recorded on a Varian Gemini-200, a Varian AMX400 or a Varian Unity Plus 500 spectrometer, operating at 200 MHz, 400 MHz, and 500 MHz for 1H NMR, respectively. Chemical shifts are denoted in δ values (ppm) relative to CDCl3 (1H: δ = 7.26; 13C: δ = 77.00). For 1H NMR, the splitting parameters are designated as follows: s (singlet), d (doublet), t (triplet), q (quartet), p (pentet), sext (sextet), m (multiplet) and b (broad). MS (EI) and HRMS (EI) spectra were obtained with a AEI
MS-137
5
BISTABLE MOTORS
902 or with a LTQ Orbitrap XL. Melting point are measured on a Büchi Melting Point B-545 apparatus. Preparative HPLC was performed on a Shimadzu semi-prep HPLC system consisting of an LC-20T pump, a DGU-20A degasser, a CBM-20A control module, a SIL-20AC autosampler, a SPD-M20A diode array detector and a FRC-10A fraction collector, using a Chiralpak (Daicel) AD-H, Chiralcel OD or Chiralcel OD-H column. Elution speed was 0.5 mL/min for AD-H and OD-H columns and 1.0 mL/min for the OD column, with mixtures of HPLC grade heptane and isopropanol (BOOM) as eluent. HPLC analysis was performed using a Shimadzu LC-10ADVP HPLC pump equipped with a Shimadzu SPDM10AVP diode array detector and chiral columns as indicated. Sample injections were made using an HP 6890 Series Autosample Injector. UV-vis absorption spectra were measured on a Analityk Jena SPECORD S600 spectrophotometer. CD spectra were measured on a Jasco J-815 CD spectrophotometer. All spectra were recorded at 20 °C using Uvasol grade heptane (Merck) as solvent. Irradiation was performed using a Spectroline ENB-280C/FE lamp (312 nm, 365 nm) or a Thorlabs INC OSL 1-EC fiber illuminator (420 nm, 450 nm). Thermal helix inversion/thermal E–Z isomerization was monitored by CD spectroscopy using the apparatus described above and a JASCO PFD-350S/350L Peltier type FDCD attachment with temperature control and cooling system or by HPLC analysis of aliquots collected over time). Temperature of oil baths during the kinetics experiments were measured with a Pt1000 RTD Temperature Sensor. Room temperature (rt) as mentioned in the experimental procedures, characterization and computational sections is to be considered equal to 20 °C.
Computational Details
Density functional theory (DFT) calculations were carried out with the Gaussian 03W (rev. C.02) program package.[72] All of the calculations were performed on systems in the gas phase using the Becke’s three parameter hybrid functional[68] with the LYP and VWN(III) correlation functionals[64,67] (B3LYP). Each geometry optimization was followed by a vibrational analysis to determine that a minimum or saddle point on the potential energy surface was found. For compounds with more than one minimum energy or saddle point conformation, the conformation with the lowest energy was chosen. For the details regarding the thermal E–Z isomerization see ref. [58].
References
[1] Chemoresponsive Materials, (Ed.: H.-J. Schneider), Royal Society Of Chemistry, Cambridge, 2015.
[2] J. Zhang, Q. Zou, H. Tian, Adv. Mater. 2013, 25, 378–99, doi:10.1002/adma.201201521.
[3] H. Dong, H. Zhu, Q. Meng, X. Gong, W. Hu, Chem. Soc. Rev. 2012, 41, 1754– 808, doi:10.1039/c1cs15205j.
[4] A. Coskun, M. Banaszak, R. D. Astumian, J. F. Stoddart, B. A. Grzybowski, Chem. Soc. Rev. 2012, 41, 19–30, doi:10.1039/C1CS15262A.
[5] B. L. Feringa, W. R. Browne, Molecular Switches Vol. I, II, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2011.
[6] W. R. Browne, B. L. Feringa, Nat. Nanotechnol. 2006, 1, 25–35, doi:10.1038/nnano.2006.45.
138
5
[7] Molecular Motors, (Ed.: M. Schliwa), Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, FRG, 2003.
[8] M. Schliwa, G. Woehlke, Nature 2003, 422, 759–765, doi:10.1038/nature01601. [9] E. R. Kay, D. A. Leigh, F. Zerbetto, Angew. Chem., Int. Ed. 2007, 46, 72–191,
doi:10.1002/anie.200504313.
[10] M. von Delius, D. A. Leigh, Chem. Soc. Rev. 2011, 40, 3656–3676, doi:10.1039/C1CS15005G.
[11] F. D. Jochum, P. Theato, Chem. Soc. Rev. 2013, 42, 7468–83, doi:10.1039/c2cs35191a.
[12] Q. Yan, Z. Luo, K. Cai, Y. Ma, D. Zhao, Chem. Soc. Rev. 2014, 43, 4199–221, doi:10.1039/c3cs60375j.
[13] M. Schwartz, Smart Materials, CRC Press, Boca Raton, USA, 2008. [14] Handbook of Stimuli-Responsive Materials, (Ed.: M.W. Urban), Wiley-VCH
Verlag GmbH & Co. KGaA, Weinheim, Germany, 2011.
[15] Intelligent Materials, (Eds.: M. Shahinpoor, H.-J. Schneider), Royal Society Of Chemistry, Cambridge, UK, 2007.
[16] M. A. C. Stuart, W. T. S. Huck, J. Genzer, M. Müller, C. Ober, M. Stamm, G. B. Sukhorukov, I. Szleifer, V. V Tsukruk, M. Urban, F. Winnik, S. Zauscher, I. Luzinov, S. Minko, Nat. Mater. 2010, 9, 101–13, doi:10.1038/nmat2614.
[17] J. Thévenot, H. Oliveira, O. Sandre, S. Lecommandoux, Chem. Soc. Rev. 2013, 42, 7099–7116, doi:10.1039/c3cs60058k.
[18] Y. Shirota, H. Kageyama, Chem. Rev. 2007, 107, 953–1010, doi:10.1021/cr050143+.
[19] Supramolecular Materials for Opto-Electronics, (Ed.: N. Koch), Royal Society Of Chemistry, Cambridge, UK, 2014.
[20] C. J. Bruns, J. F. Stoddart, Acc. Chem. Res. 2014, 47, 2186–2199, doi:10.1021/ar500138u.
[21] G. Du, E. Moulin, N. Jouault, E. Buhler, N. Giuseppone, Angew. Chem., Int. Ed.
2012, 51, 12504–12508, doi:10.1002/anie.201206571.
[22] A. Coskun, J. M. Spruell, G. Barin, W. R. Dichtel, A. H. Flood, Y. Y. Botros, J. F. Stoddart, Chem. Soc. Rev. 2012, 41, 4827–59, doi:10.1039/c2cs35053j.
[23] T. Kudernac, N. Ruangsupapichat, M. Parschau, B. Maciá, N. Katsonis, S. R. Harutyunyan, K.-H. Ernst, B. L. Feringa, Nature 2011, 479, 208–11, doi:10.1038/nature10587.
[24] J. Wang, B. L. Feringa, Science 2011, 331, 1429–1432, doi:10.1126/science.1199844.
[25] S. J. Wezenberg, M. Vlatković, J. C. M. Kistemaker, B. L. Feringa, J. Am. Chem. Soc. 2014, 136, 16784–16787, doi:10.1021/ja510700j.
[26] W. M. Horspool, F. Lenci, CRC Handbook of Organic Photochemistry and Photobiology, Vol. 1 & 2, CRC Press, Boca Raton, USA, 2003.
[27] G. S. Kottas, L. I. Clarke, D. Horinek, J. Michl, Chem. Rev. 2005, 105, 1281–376, doi:10.1021/cr0300993.
[28] B. L. Feringa, J. Org. Chem. 2007, 72, 6635–52, doi:10.1021/jo070394d. [29] N. Koumura, R. W. Zijlstra, R. A. van Delden, N. Harada, B. L. Feringa, Nature
139
5
BISTABLE MOTORS
[30] N. Koumura, E. M. Geertsema, M. B. van Gelder, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2002, 124, 5037–5051, doi:10.1021/ja012499i.
[31] J. C. M. Kistemaker, P. Štacko, J. Visser, B. L. B. L. Feringa, Nat. Chem. 2015, 7, 890–896, doi:10.1038/nchem.2362.
[32] E. M. Geertsema, S. J. van der Molen, M. Martens, B. L. Feringa, Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 16919–24, doi:10.1073/pnas.0903710106.
[33] R. A. van Delden, M. K. J. ter Wiel, H. de Jong, A. Meetsma, B. L. Feringa, Org. Biomol. Chem. 2004, 2, 1531–41, doi:10.1039/b402222j.
[34] J. Vicario, A. Meetsma, B. L. Feringa, Chem. Commun. 2005, 5910–2, doi:10.1039/b507264f.
[35] D. Pijper, R. A. van Delden, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2005, 127, 17612–3, doi:10.1021/ja054499e.
[36] J. Vicario, M. Walko, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc. 2006, 128, 5127–35, doi:10.1021/ja058303m.
[37] M. M. Pollard, M. Lubomska, P. Rudolf, B. L. Feringa, Angew. Chem., Int. Ed.
2007, 46, 1278–80, doi:10.1002/anie.200603618.
[38] M. Klok, N. Boyle, M. T. Pryce, A. Meetsma, W. R. Browne, B. L. Feringa, J. Am. Chem. Soc. 2008, 130, 10484–5, doi:10.1021/ja8037245.
[39] A. Cnossen, D. Pijper, T. Kudernac, M. M. Pollard, N. Katsonis, B. L. Feringa, Chem. - Eur. J. 2009, 15, 2768–72, doi:10.1002/chem.200802718.
[40] T. C. Pijper, D. Pijper, M. M. Pollard, F. Dumur, S. G. Davey, A. Meetsma, B. L. Feringa, J. Org. Chem. 2010, 75, 825–38, doi:10.1021/jo902348u.
[41] A. A. Kulago, E. M. Mes, M. Klok, A. Meetsma, A. M. Brouwer, B. L. Feringa, J. Org. Chem. 2010, 75, 666–79, doi:10.1021/jo902207x.
[42] L. F. Tietze, M. A. Düfert, T. Hungerland, K. Oum, T. Lenzer, Chem. - Eur. J.
2011, 17, 8452–8461, doi:10.1002/chem.201003559.
[43] L. Greb, J.-M. Lehn, J. Am. Chem. Soc. 2014, 136, 13114–13117, doi:10.1021/ja506034n.
[44] J. Bauer, L. Hou, J. C. M. Kistemaker, B. L. Feringa, J. Org. Chem. 2014, 79, 4446–4455, doi:10.1021/jo500411z.
[45] M. K. J. ter Wiel, R. A. van Delden, A. Meetsma, B. L. Feringa, J. Am. Chem. Soc.
2005, 127, 14208–22, doi:10.1021/ja052201e.
[46] R. Göstl, A. Senf, S. Hecht, Chem. Soc. Rev. 2014, 43, 1982–96, doi:10.1039/c3cs60383k.
[47] V. Blanco, D. A. Leigh, V. Marcos, Chem. Soc. Rev. 2015, 44, 5341–5370, doi:10.1039/C5CS00096C.
[48] M. Klok, M. Walko, E. M. Geertsema, N. Ruangsupapichat, J. C. M. Kistemaker, A. Meetsma, B. L. Feringa, Chem. - Eur. J. 2008, 14, 11183–93,
doi:10.1002/chem.200800969.
[49] G. B. Kistiakowsky, W. R. Smith, J. Am. Chem. Soc. 1934, 56, 638–642, doi:10.1021/ja01318a029.
[50] A. V. Santoro, E. J. Barrett, H. W. Hoyer, J. Am. Chem. Soc. 1967, 89, 4545–4546, doi:10.1021/ja00993a066.
[51] I. Agranat, M. Rabinovitz, A. Weitzen-Dagan, I. Gosnay, J. Chem. Soc. Chem. Commun. 1972, 732–733, doi:10.1039/C39720000732.
140
5
[52] P. U. Biedermann, J. J. Stezowski, I. Agranat, European J. Org. Chem. 2001, 2001, 15–34,
doi:10.1002/1099-0690(200101)2001:1<15::AID-EJOC15>3.0.CO;2-0.
[53] A. Kazaryan, J. C. M. Kistemaker, L. V. Schäfer, W. R. Browne, B. L. Feringa, M. Filatov, J. Phys. Chem. A 2010, 114, 5058–5067, doi:10.1021/jp100609m. [54] G. Pérez-Hernández, L. González, Phys. Chem. Chem. Phys. 2010, 12, 12279–89,
doi:10.1039/c0cp00324g.
[55] A. Cnossen, J. C. M. Kistemaker, T. Kojima, B. L. Feringa, J. Org. Chem. 2014, 79, 927–935, doi:10.1021/jo402301j.
[56] B. O. Roos, P. R. Taylor, P. E. M. Siegbahn, Chem. Phys. 1980, 48, 157–173, doi:10.1016/0301-0104(80)80045-0.
[57] It is noted that multi-reference methods based on the density matrix
renormalization group (DMRG) approach are capable of handling much larger active spaces. Unfortunately, however, this approach is unavailable in almost all QC software packages.
[58] Supplementary Material is available online as a part of this manuscript (doi:10.1002/chem.201602276), can be found under
http://onlinelibrary.wiley.com/store/10.1002/chem.201602276/asset/supinfo/chem2
01602276-sup-0001-misc_information.pdf?v=1&s=6933b84dd7390a09837e57bd64c718be3100b3ab. [59] Comprehensive Chiroptical Spectroscopy, (Eds.: N. Berova, P.L. Polavarapu, K. Nakanishi, R.W. Woody), John Wiley & Sons, Inc., Hoboken, NJ, USA, 2012. [60] The Lambert W function solves y=x.e^x for x as x=W(y), see ref. [31]. [61] W. J. Hehre, J. Chem. Phys. 1972, 56, 2257–2261, doi:10.1063/1.1677527. [62] P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973, 28, 213–222,
doi:10.1007/BF00533485.
[63] J. S. Binkley, J. A. Pople, W. J. Hehre, J. Am. Chem. Soc. 1980, 102, 939–947, doi:10.1021/ja00374a017.
[64] S. H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 1980, 58, 1200–1211, doi:10.1139/p80-159.
[65] M. M. Francl, J. Chem. Phys. 1982, 77, 3654–3665, doi:10.1063/1.444267. [66] M. S. Gordon, J. S. Binkley, J. A. Pople, W. J. Pietro, W. J. Hehre, J. Am. Chem.
Soc. 1982, 104, 2797–2803, doi:10.1021/ja00374a017. [67] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789,
doi:10.1103/PhysRevB.37.785.
[68] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652, doi:10.1063/1.464913. [69] T. C. Pijper, Optically Responsive Switches [Groningen], PhD Thesis, University
of Groningen, 2015.
[70] J. C. M. Kistemaker, S. F. Pizzolato, T. van Leeuwen, T. C. Pijper, B. L. Feringa, Chem. - Eur. J. 2016, 22, 13478–13487, doi:10.1002/chem.201602276.
[71] S. F. Pizzolato, Dynamic Transfer of Chirality in Photoresponsive Systems [Groningen], PhD Thesis, University of Groningen, 2017.
141
5
BISTABLE MOTORS
[72] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09, Revision C.01, Gaussian, Inc., Wallingford CT, 2009.
